The Borrowing Puzzle: Why Do FilipinoDomestic Workers in Hong Kong Borrow

rather than Dissave?a

Wooyoung Limb Sujata Visaria c

March 2, 2020

Abstract

Despite their predictable and regular incomes, Filipino domestic workersin Hong Kong commonly finance large expenses through interest-bearingloans rather than savings. Our analysis of survey data and the records ofa credit cooperative for migrant workers suggests that this cannot be ex-plained by their inability to save, by financial illiteracy, a short time horizonor limited liability. Instead, we speculate that the strict schedules and highinterest rates of these loans create a disciplining effect that these individ-uals find desirable. This may help them avoid unnecessary consumption,or demands from their social network. However interventions should alsoconsider the fact that members of the social network often provide non-monetary reciprocal benefits.Keywords: migrants, savings, loans

aA large team of HKUST UROP students, ably led and supported by Arpita Khanna, Sheren Ku and Jimmy Santiagohelped collect the data for this paper. The Asian Migrants Credit Union kindly shared their records. Ethics approval wasobtained from HKUST. We thank Rina Lookman Jio and Ziyi Hong for their terrific help analysing the data, UtpalBhattacharya, Clarence Lee, Dilip Mookherjee, Jane Y. Zhang and seminar audiences at the 2019 Asian DevelopmentBank RoundTable and the Indian Institute of Management Ahmedabad for insightful conversations and comments.This research was funded by an HKUST Institute for Emerging Market Studies Research Grant.

bDepartment of Economics, Lee Shau Kee Business Building, Hong Kong University of Science and Technology,Clearwater Bay, Hong Kong. wooyoung@ust.hk.

cDepartment of Economics, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong. svis-aria@ust.hk.

1 Introduction

Domestic workers make up a significant flow of migrants in Asia. Employersfrom several higher-income countries such as Hong Kong, Malaysia and Singa-pore recruit live-in domestic help from countries such as the Philippines andIndonesia. This migration is temporary and is almost entirely motivated by eco-nomic gain. However, our understanding of the benefits of such migration onthe migrant’s household is sparse. Migrant remittances have been shown to im-prove the contemporaneous living standards and educational investments ofdependents, but what are the long-term economic gains and opportunities forupward mobility that such migration enables? In particular, are these migrantseconomically self-sufficient in their retirement after they return to their homecountries? Does the migration successfully enable a transition to higher-wageoccupations for the next generation?

Definitive answers to these questions require large-scale and long-term datacollection and plausibly exogenous variation in the migration decision. How-ever even in their absence, we can observe and analyse the constraints andchoices among migrants, and make inroads toward an understanding. In thispaper we begin by acknowledging that migrants’ financial choices during theirtenure in the host country greatly influence their and their households’ out-comes in the future. Whether their incomes are spent only on consumption orare also saved, and whether they make productive investments for financialgain, will determine both whether they will retire comfortably and also whetherthe migration will improve their economic status.

Our particular study population is Filipino domestic workers in Hong Kong.We examine how they manage their finances, specifically, their choice betweensavings and loans. As we will document, it would appear that their financialchoices do not maximise their economic gains. We find that they commonly fi-nance foreseen, discretionary investments through debt rather than savings –taking interest-bearing loans from moneylending companies rather than firstbuilding up their savings and then dissaving cheaply. In fact, we find evidenceof “co-holding”, i.e. that they hold liquid savings and borrowing at the sametime. We will argue that this imposes a significant financial cost and yet offers

1

no financial benefit: the debt contracts are not designed to transfer project riskto the lender, interest rates are non-negligible, and in fact, loan default imposesa heavy cost, with the real risk of losing their job and cutting off future earningsin Hong Kong.

To identify and then examine this “borrowing puzzle”, we draw on datacollected through a survey, a lab-in-the-field experiment, and the records of acredit cooperative that caters to migrant workers in Hong Kong. In 2017, we in-terviewed a sample of 136 Filipino domestic workers and asked about their em-ployment history, wage income, remittances, savings, and loans. Subjects alsoparticipated in a lab-in-the-field experiment where they allocated a given en-dowment among a set of options with differing risk and returns.

We document the following facts. One, as expected, most Filipino domesticworkers remit to their home country regularly. Often these remittances supportnot just their immediate nuclear family, but support educational and health ex-penses for their extended family as well. Thus these migrants appear to take onthe responsibility of supporting several individuals back home.

Two, although the majority of migrants have bank accounts, they do not ap-pear to use them as a savings device. Bank balances tend to be low and monthlyinflows into the accounts are small. However this is not to say that their entiremonthly salary is consumed. Anecdotal evidence suggests that many of theminvest in “projects” in the Philippines, such as land purchase, house construc-tion, house renovation and repair, and small businesses.

Three, it is common for them to borrow from moneylending companies inHong Kong. On average these companies charge 25 percent interest per annum.The migrants repay these loans from their salaries in Hong Kong. Our data sug-gest that only a small fraction of these loans are used for unforeseen emergencyexpenses; the majority are remitted home for school fees, consumption needs orfor investment.

This leads us to the central observation in this paper: Filipino domestic work-ers appear to routinely finance their investments through loans rather than sav-ing. Moneylending companies have standard contracts for loans to overseasworkers, where repayment begins the very next month after the loan is dis-

2

bursed. The investments generally do not start generating immediate returns,so that we argue that the repayment is financed from the worker’s wages inHong Kong. These wages are contracted and regular, and therefore predictable.Default carries heavy penalties, thus borrowing does not transfer risk to thelender. We consider different explanations for why many migrants would ratherfinance investments through borrowing than dissaving. Our data allow us toexamine the plausibility of these different explanations. Although we cannotconclusively accept or reject a particular hypothesis, we discuss possibilities forfuture research that could shed light on this issue.

2 The Context

Migrant domestic workers made up 9.3% of Hong Kong’s workforce in 2016. More than half of these were from the Philip-pines[Government of the Hong Kong SAR, 2017]. They perform a range ofservices for their employers, including cleaning, cooking, shopping for gro-ceries, babysitting, ferrying children to and from school and extra-curricularactivities, and caring for the employers’ aged parents and pets. Their servicesfacilitate the labour force participation of working-age Hong Kong women,especially mothers with young children[Cortes and Pan, 2013].1

The Philippines was one of the first countries to send workers through HongKong’s foreign domestic helper program that began in the 1970s. This programgrants migrants a special “foreign domestic helper” visa, which entitles themto work for a single employer in Hong Kong. Employers are required to payat least a “minimum allowable wage”. As of 2020 this minimum wage is HKD4630; it is usually revised once a year to adjust for changes in the cost of liv-ing.2 Foreign domestic helpers cannot qualify for permanent residence in HongKong. They can continue to reside in Hong Kong as long as they are gainfullyemployed as domestic workers, but the employment contract and visa must be

1The benefits from employing foreign domestic workers likely extend beyond the effect onfemale labour supply. Tan and Gibson (2013) argue that domestic workers do not increase fe-male labour force participation in Malaysia, but speculate that Malaysian employers can enjoyincreased leisure and can specialize in childrearing, with possible gains in their childrens humancapital.

2Note that by law the worker must live in the residence of the employer, and so she is notexpected to incur any housing costs.

3

renewed every two years.3 Thus these individuals are temporary economic mi-grants: they live in Hong Kong only for as long as they can be gainfully legallyemployed. They are generally aware that as they get older, employers becomeless likely to employ them, and that they will retire in their home country.

These migrants’ wages are lower than most of the Hong Kong population.4

They do not qualify for pensions or other financial benefits. Commercial banksusually target higher-income groups, as a result foreign domestic workers haveonly limited access to formal banking services. However Hong Kong’s laws putno restrictions on whether and how much migrants can borrow from local mon-eylending companies.5 This creates an interesting dichotomy where they haveonly limited access to formal savings accounts, but extensive access to formalloans. This paper examines how they manage their finances against this back-drop.

3 Data Collection

In 2017, we enrolled 141 Filipino domestic workers to participate in our surveyand lab-in-the-field experiment. Of these, we have survey data from the 136 whosuccessfully completed the two parts of our interview.6 Below we describe theprocess by which this sample was created.

Migrant domestic workers in Hong Kong are required by law to live in theiremployer’s house. They usually work 6 days per week for unspecified hours,and then spend most of their weekly holiday outside the employer’s house.

3If they become unemployed, migrant domestic workers must leave Hong Kong within twoweeks and may only re-enter after an employer has signed a new two-year contract. Note thatwhen they fire a domestic worker, employers are required to pay for the worker’s travel back totheir home country.

4There are no regulations about or definitive data on the number of hours that they work perday. However taking a conservative estimate of 10 hours per day for 6 days a week, the monthlyminimum allowable wage for domestic helpers in 2019 translates to an hourly rate just belowHK$20, or roughly 50% of the Hong Kong minimum wage. Domestic helpers do not fall underthe purview of the minimum wage ordinance.

5In contrast, in 2019 the Singapore government placed a limit on how much individuals canborrow from Singaporean moneylenders. Some have even argued that domestic workers’ loanapplications should be pre-approved by their employers[Ng and Tan, 2019].

6All 141 participated in the first face-to-face interview, but only 136 could be contacted 4-8weeks later for a phone interview that asked about financial transactions that had occurred sincethe first interview, and asked additional questions that helped to compute loan interest rates.

4

This makes it very difficult for investigators to survey them in their residence.Given the length of our survey and experimental sessions, we believed it wouldbe difficult to enroll participants by approaching them on the street or in pub-lic places.7 We therefore advertised our study through Whatsapp and Facebookwith certain nodal Filipino domestic workers, and asked them pass the adver-tisement on.8 Interested persons could click on a web link and answer a short en-rollment questionnaire. We then contacted these enrollees and reserved a studysession slot for them.9 In this way we created a respondent-driven sample.

A respondent-driven sample may not be representative of the underlyingpopulation of interest. The nodes that we began with were not randomly cho-sen, and if we had only relied on the nodes to spread the word, we might haveonly reached their friends. We therefore attempted to create a “snowball” byoffering each participant a bonus per referred person who also signed up forour study. This created an incentive for every participant to spread the wordto her friends. Again, to avoid swamping the sample with the acquaintancesof the more popular participants, we offered this bonus for only 4 referrals,and no more. However, snowball samples only approximate random samplesin the limit, and our sample of 136 respondents is unlikely to be sufficientlylarge [McKenzie and Mistiaen, 2009]. We therefore present re-weighted descrip-tive statistics that better approximate the true population of Filipino domesticworkers living in Hong Kong.10

We also draw on the records of the Asian Migrants Credit Union, a HongKong credit cooperative that primarily targets migrant domestic workers. All in-dividuals who join the cooperative are provided a savings account. Six monthsafter they enroll, members become eligible to use the credit facility. First-time

7Relatedly, Barua, Shastry, and Yang [2019] found that domestic workers recruited in publicplaces in Singapore were unlikely to sustain participation in their study.

8These nodal persons were elected officers of the credit cooperative that we also use datafrom.

9Nearly all migrant domestic workers in Hong Kong use smartphones, and a very large frac-tion use social media, and so it was fairly easy for them to view and answer our enrollmentquestionnaire. Once we received their online submission we called them to explain the detailsof the study session and register them into a time slot.

10We draw the weights from the distribution of Filipino domestic workers’ age, educationlevel and length of stay in the 5% micro-sample of the 2016 Hong Kong By-census. The Ap-pendix provides further information about this micro-sample and how we use it.

5

borrowers may only borrow up to two times their savings. The entire loan iscollateralised by the savings balance of the borrowing member as well as thesavings of the guarantors, who are other members of the cooperative. The inter-est rate is set at 1 percent per month and repayment is on a monthly schedule.We look at the contracts of all loans that were approved by the cooperative in2017-18, to examine the stated purposes of these loans. We also analyse mem-bers’ choices between saving and borrowing during the period 2011-18.11

Finally, we also draw on the findings of a lab-in-the-field experiment with all141 subjects who participated in our study. The goal of the study was to examinehow participants respond to the rate of return when making savings choices.All subjects were given an endowment of 100 tokens (1 token was equivalentto $1), and in each round they were asked to allocate these tokens across threeaccounts: a savings account that generated a sure return, an investment accountthat would generate a return of 10 percent but with uncertainty, and a lotteryaccount where each token would give them a chance of winning a handbag as aprize. Subjects were randomly assigned to groups and played multiple roundswithin each group.12

4 Some Facts

Table 1 presents re-weighted descriptive statistics about the respondents in oursurvey. The average Filipino domestic worker was 36.5 years old. She had leftthe Philippines for work about 6.5 years prior to our study, had been in HongKong for nearly 5 of those years and had worked for her current employer forabout 3 years. Given their profession, domestic workers reported relatively higheducation levels. A third had studied beyond high school, and a fifth had anadditional academic qualification beyond high school.13

As we mentioned before, migrant domestic workers must work for onlyone employer. Contracts are signed for a 2-year duration. Our study took place

11Although we have the transaction records for the period 2011-18 for the credit union, weonly have access to the loan applications for 2017-18.

12Further details about this study are provided in the appendix.13This need not be a university degree; the category includes associate degrees, vocational

training and professional courses.

6

between January and May 2017, thus the workers we interviewed must havesigned their current contracts no sooner than May 2015. The minimum allow-able wage was set at HKD 4310 for contracts signed between October 2016 andSeptember 2017, and HKD 4210 for contracts signed between October 2015 andSeptember 2016. The median worker in our sample received exactly HKD 4210,but the mean was a slightly lower HKD 4150. Overall, the evidence suggests thatemployers comply with the minimum allowable wage regulation. More than 80percent of workers received the wages as cash, which is indicative of the lowuse of formal banking or other financial services.

As is to be expected, a large party of their salaries were remitted back to theirhouseholds in the Philippines. Eighty-eight percent of the sample had remittedmoney home within the last two months. On average, they remitted HKD 2164,or 52 percent of their monthly salary. These remittances supported on average3.7 individuals. This included not just her immediate family but the extendedfamily as well.14

4.1 Savings

Strikingly, 83 percent of our sample had active bank accounts at the time of thesurvey. Of these, 91 percent of workers had accounts in the Philippines, whereasa much lower 32 percent had accounts in Hong Kong. Nearly all bank accountswere single-holder accounts; only 6 percent of account holders reported hav-ing joint accounts. However bank balances were low. Across both Hong Kongand the Philippines, the average respondent held only about HKD 5840 or 1.4months’ salary in the bank. Net inflows were actually negative during the periodthat we asked about, viz. the 2 months prior to the interview. Other savings de-vices were not very common. Only 11 percent reported membership in a Rosca,where they made a monthly contribution of HKD 340 on average. Nobody re-ported using a money-guarding arrangement.15

14Fifty-six percent were supporting their parents, and 34 percent were supporting other de-pendents, such as grandchildren, siblings, nieces and nephews, cousins or grandparents.

15A money-guard is a person who holds money for the subject to help her avoid spending orlosing it [Collins, Rutherford, Morduch, and Ruthven, 2009].

7

4.1.1 Savings Response to Rates of Return

Migrant workers may have held small bank balances because the accounts of-fered low rates of return.16 To examine whether migrants’ savings balances re-spond to interest rates, our experiment randomly assigned participants to a“savings product” with one of two rates of return: a low 3% rate, or a high 10%rate. Strikingly, we find no evidence that participants assigned to the high returncondition allocated more tokens into the safe account. Respondents in the lowreturn condition placed $53.3 out of $100 worth of tokens into the safe account,and those in the high return condition placed a nearly identical $51.7 (difference= 1.6, p=0.68). The rate of return on savings also did not affect the allocation tothe other two accounts.17

To understand whether this behaviour can be explained by migrant charac-teristics, in Table 2 we examine whether respondents with different characteris-tics respond differently to the change in the rate of return. Our data consist of 324person-round level observations across the 141 respondents who participatedin the experiment. Our regressions includes dummy variables for the round inwhich the allocation was made. This controls for round-specific effects, or learn-ing over time. In column (1) we include as an explanatory variable a measureof the respondent’s risk aversion.18 As expected, we find that more risk averserespondents placed a larger number of tokens in the safe return box. In column(2) we include controls for age, education and length of stay in Hong Kong;this does not change the coefficient on risk-aversion significantly. In column (3),we add a dummy variable for whether the respondent faced the 10% returnon savings. Controlling for risk-aversion, we do not find that respondents whofaced a higher rate of return placed more tokens than comparable respondentswith a lower rate of return. Finally, in column (4) we interact the risk-aversion

16Hong Kong savings interest rates are nearly 0%. At 0.1%, interest rates in the Philippinesare only slightly higher. Inflation rates in 2017 were 1.48% in Hong Kong, and 2.85% in thePhilippines.

17Investment account: $32.2 in the low savings return condition v. $30.5 in the high returncondition (difference = 1.7, p=0.64); Lottery account: $23.3 v. $24.4 (difference = 1.1, p=0.73).

18Risk preferences were elicited using an incentivised Lowry list method where participantswere asked to choose between a safe option and a lottery with a high and a low payout, wherethe probability of the high payout successively increased. In line with Yu, Zhang, and Zuo[2019], respondents were encouraged to choose a single switching point from safe option tolottery.

8

measure with the dummy variable for the high-return treatment. There is noevidence that more risk-averse individuals responded differently to the rate ofreturn than the less risk-averse. In columns (5)-(8) we consider heterogeneouseffects by the respondent’s financial literacy.19 Again, there is no evidence thatfinancial literacy levels affected how participants responded to the rate of re-turn. To the extent that these results can be translated into their behaviour indaily life, it does not appear that migrants’ disinterest in saving is driven by thelow rates of return.20

Note also that the credit cooperative paid considerably higher dividends (1to 3% per annum) than the Hong Kong commercial bank interest rate duringthis period. Despite this, we find a low savings rate among members of thecredit cooperative. The average member made a net deposit of only HKD 44per month into her account. There is also no evidence that the members’ sav-ings rates responded to dividends. In Figure 1 we plot the monthly net depositsper member against the dividend rate that the credit union paid in the previ-ous year.21 There is no indication that members saved more per month whendividends were higher.

4.2 Credit

Loans allow individuals with small cash inflows to consume or invest in thepresent, instead of having to postpone or forego these activities. They can alsohelp smooth consumption in the face of negative shocks. Below we examine thenature of borrowing by our sample subjects.

Once again, our summary statistics have been re-weighted to match the dis-tribution of Filipino domestic workers in the 2016 Hong Kong by-census. Forty-six respondents (or 37 percent of our re-weighted sample) reported that at thetime of the survey they had an outstanding loan from a lender in Hong Kong.

19In Section 5.3 we describe how we measured financial literacy.20As we see across the table, respondents with a post-high-school qualification place more

tokens in the safe return box. However there is no evidence that they increase their safe tokenallocation when they face a higher safe return (results available upon request).

21Each year’s dividends are announced at the end of the year and depend on the credit union’sprofits during the year. Arguably members could not have known the dividend when they madethe saving decision, but they could have used the previous year’s dividends as a predictor.

9

When we compute their monthly repayment obligation we find that on aver-age they had committed to paying 55 percent of their salary in loan installmentseach month. Table 3 also shows that 88 percent of the loans had been takenfrom moneylending companies. The cooperative only gave out 3 percent of theloans. The other sources were informal: either informal borrowing from theiremployers, or loans from friends and relatives in Hong Kong. Loans from mon-eylending companies were the largest of all: the average principal amount wasHKD 22176. Loans were for an 11-month duration on average, the interest ratewas 26 percent per annum, and payment was due on a monthly basis.

Employers gave zero-interest loans. An average loan given by an employerwas for HKD 15,339 on average, or just over 3 months’ salary. Employers alsotook payment on a monthly basis, usually by garnishing the worker’s salarypayment. Loans from the credit cooperative were similar in size at HKD 13154on average. At 7 percent per annum, the cooperative charged less than one-thirdthe interest of the moneylending companies.22 Loans from friends and relativeswere significantly smaller, and although many subjects expected to repay on amonthly basis, the repayment schedules were more likely to be flexible.

In keeping with our finding that respondents had received loans fromfriends and relatives in Hong Kong, 14 percent of our survey respondents toldus that either friends or relatives in Hong Kong owed them money at the cur-rent time. Of the 27 such loans, 5 represented a sharing arrangement where therespondent had taken a loan from a moneylending company and then shared itwith another domestic worker.23 Another 3 were given out by a single respon-dent to three different friends at 10 percent interest over a 6 month duration, or1.67 percent per month.

22The cooperative capped the loan principal at 2 times the borrowing member’s savings bal-ance in the cooperative.

23This is an informal arrangement where the individual who is formally listed as the borrower“shares” some of the loan principal with a friend. Often this friend is either the reference personor the guarantor for this loan. Both friends then pool together to make up the monthly payments.In case of default the moneylender first contacts the borrower and then contacts the referenceperson/guarantor. They may also contact the employer of either or both domestic workers anddemand payment from them. It is possible that this informally-created joint liability improvesloan repayment for both the borrower and the sharer. Moneylenders offer “VIP status” to bor-rowers with good repayment records. VIPs earn in-kind rewards and rebates on their own loanpayments in return for referring new borrowers to the moneylending company.

10

Thus the survey data suggest that it is common for Filipino domestic workersto borrow. In fact, the records of the credit cooperative suggest that migrantsneed not borrow as much as they do. We present below evidence that a largeproportion of cooperative members “co-hold” loans at the same time as theyhold liquid savings that could be drawn down instead.

Since the interest cost on these loans (1 percent per month) is considerablyhigher than the return on savings (AMCU dividend rates range from 1 to 3%per year), it is in the members’ interest to take the smallest loan necessary tofinance their need. However, in 17.5% of the 200 loans that the cooperative gaveout over the period 2011-2017, the member had larger savings in the cooperativethan the amount she borrowed. Clearly it would have been cheaper to insteadwithdraw these savings and avoid the loan altogether. Instead, by taking theloan and securing it with part of her savings, she both took on an additionalinterest expense, and rendered part of her savings illiquid.24

Second, even when the member’s savings were smaller than her loanamount, the evidence suggests she could have borrowed less than she did.Specifically, she could have withdrawn part of her savings, thereby reducingthe loan size and interest cost, while still financing the expense. To see this, notethat one-half of the loan is secured by the member’s savings and therefore can-not be withdrawn. Call this her illiquid savings, i. The remaining savings isliquid; denote this by l. Thus total savings s = i + l. If the expense is e then weknow that e − l = 2i. We can then calculate her illiquid savings as i = e − s,so that the remaining l = 2s − e can be withdrawn. By withdrawing the entirel, a member would take the smallest loan necessary to finance the expense. Forexample, a member who needs to finance an expense of $1000 and has $700 insavings would minimise costs by maintaining a savings balance of $1000 - 700= $300 to take a loan of $600, and withdrawing the remaining $400.

Instead, we find that in 62.5 percent of the credit cooperative loans, the loanprincipal was less than two times the member’s savings balance at the time.25

24The cooperative limits a member’s loan amount to twice her total savings balance at thetime of the loan application. This savings balance secures the loan and cannot be withdrawnwhile the loan is outstanding.

25Our methodology and findings are similar to that used by Laureti [2018] in her analysis ofthe clients of SafeSave, which provides flexible savings-and-loans accounts to slum-dwellers in

11

The data thus suggest that it is common for Filipino domestic workers to fi-nance their investments and perhaps even their families’ consumption expensesthrough loans, rather than saving. In what follows, we consider different expla-nations for why they may do so.26

5 Explanations for Indebtedness

We start by examining some common explanations for the high incidence ofdebt.

5.1 Debt due to Migration Costs

It is widely reported in Hong Kong that domestic workers bear alarge financial cost to get a job placement. This appears to applyboth to workers located in the Philippines looking to migrate, as wellas workers located in Hong Kong who are in between employers.PLUDWHK and Hong Kong Federation of Asian Domestic Workers [2016] re-ports on an investigation where researchers made anonymous phonecalls pre-tending to be domestic workers in search of employment. They found that mostemployment agencies charge workers a sizeable illegal fee for the placementservice. If the worker is unable to pay the placement fee upfront, the employ-ment agency often refers her to a lending company. Job applicants can also takeloans in the Philippines and repay them from Hong Kong. It is reported that theaverage Filipino domestic worker takes 6 months to pay off this loan.

Thus, migrants might be arriving in Hong Kong already in debt, and maycontinue to be indebted for a significant duration of their first contract. If theyincur a placement fee again when they switch employers, then they may need totake another loan and could be indebted for part of their first contract with thenew employer. If they faced any large unexpected consumption or investment

Dhaka, Bangladesh.26When a cooperative member over-borrows, she takes a larger loan than necessary, but also

secures a larger fraction of the loan with her own savings, thereby relying on a guarantor tosecure a smaller fraction. If she instead withdrew her savings and took a smaller loan she wouldstill need a guarantor to secure the exact same dollar amount. Thus over-borrowing does notreduce dependence on the guarantor.

12

expense during this period, they may need to take another loan, possibly settingthem on a path of repeated indebtedness.27

If indebtedness is explained by the placement fee expense, then we shouldsee greater indebtedness among migrants who are in their first contracts, thanmigrants who have been with their current employer for longer.28 In fact, only29% of our sample who were in their first contract currently had a loan in HongKong, versus 37% of those who had been with their current employer for longer(difference statistically non-significant).29

5.2 Unexpected Expenses

We have referred above to the possibility that negative shocks may induce mi-grants to borrow to smooth consumption. If migrants are using the bulk of theirincomes to support the regular expenses of their families, then even a migrantwho saves regularly may simply not have enough saved up to cope with ashock. However, the data from the credit cooperative suggest that this cannot bea complete explanation. When we analyse the stated purpose of the 40 loans thatmigrant domestic workers took from the credit cooperative in the year 2017-18,we find that nearly two-thirds were for expenses that could have been antici-pated: land purchase, home renovation, or school fees for children back home.Only 21% of loans were for the medical expenses of relatives.30 Of course, thereis the question about whether we can trust the stated purpose of the loan. How-ever the credit cooperative has an informal policy of providing faster customer

27Recently, three major moneylending companies appear to have started sharing informationabout their clients’ loan records; in informal conversations domestic workers report that theycan no longer take multiple loans from different moneylenders.

28Recontracting with the same employer is a relatively easy process, and most domestic work-ers do not rely on employment agencies for it.

29The low incidence of loans among those in their first contract may seem puzzling. Notingthat it is illegal to charge placement fees to workers, it is possible that more experienced workersare more aware of this rule, or can find employment more easily through word-of-mouth orother means, rather than using an agency. We therefore test, but reject the hypothesis that newarrivals to Hong Kong are more likely to be in debt. This could be because their loans were takenin the Philippines and are therefore are not being reported as Hong Kong loans. In any case, thisdoes not suggest that migration-related costs are causing the indebtedness we observe in HongKong.

30The rest could not be cleanly classified into emergency or non-emergency purposes. Forexample, house repair could be an urgent expense in response to sudden damage, or could be anon-urgent expense that was planned ahead.

13

service for emergency loans, and so it seems unlikely that borrowers under-report the true incidence of emergencies. Rather, it appears that the bulk of theloans are not being used to smooth shocks.

5.3 Lack of Financial Knowledge

Domestic workers may not understand the financial cost of borrowing. In otherwords, they may not realise that they can lower their financial costs by usingtheir savings instead of borrowing. To examine whether this can explain theobserved behaviour, we examine whether indebtedness varies by financial liter-acy levels. Our measure of financial literacy comes from two questions we askedin our survey. In each question, the respondent was presented with two alter-native hypothetical loans and needed to determine which loan was cheaper.31

The respondent’s financial literacy score takes a value of 0, 1 or 2 depending onwhether she correctly identified the cheaper loan in none, one, or both of thequestions.

First, our data do not suggest that respondents are generally unable to evalu-ate the cost of a loan. Fifty-one percent of respondents answered both questionscorrectly, and 40 percent answered one correctly. However 37% of those whoanswered both questions correctly reported having an outstanding loan, com-pared to 30% of the others. This difference is not statistically significant. Thusit does not appear that their behaviour stems from an inability to compute thefinancial costs.

5.4 Lack of Self-Control or Other-Control

Hong Kong is a consumerist society, and shopping opportunities are every-where. It could be argued that this creates the temptation for Filipino migrantsto purchase goods that may not be strictly necessary. Excessive consumption of

31In question 1, the two loans had an identical duration but differed in the principal and theinterest. Thus respondents would have needed to work out which loan was cheaper per dollarof principal. In the second question, the two loans had identical principal and duration, but theloan installment size and installment frequency varied. Thus they would have needed to workout which loan required the larger repayment amount. The exact questions are reproduced inthe Appendix. For each pair of loans we asked them two questions: which loan was cheaper,and which loan they would prefer to take.

14

these goods could prevent them from building up their savings, so that theymight need to borrow to finance large expenses.

Alternatively, migrants could lack complete property rights over their earn-ings. In other words, they could be remitting larger sums than they had plannedto, or purchasing items that they did not plan to, not because they lack self-control but because others in their social network make demands on their in-comes. For example, their families back home may demand gifts, or ask forlarger remittances. Similarly if they have surplus cash, their friends in HongKong may request loans or treats and these may not be repaid or reciprocated.

Either of these two mechanisms could lead them to have low savings, neces-sitating that they borrow to finance large expenses. Although our current datado not allow us to validate these explanations, we will discuss these mecha-nisms further in Section6.3.

6 Explanations for Over-borrowing

As we discussed above, the puzzle is not only that savings tend to be low onaverage, but that often individuals choose to borrow, instead of withdrawingtheir savings. We now discuss some explanations for this behaviour.

6.1 Limited Liability

Recall that the stated purpose for most credit cooperative loans was often aproductive investment such an educational expense, property purchase or con-struction, or a business investment. These are potentially risky investments. Ifloan contracts offer borrowers limited liability, then by financing the investmentthrough a loan, the migrant worker transfers the downside risk to the lenderand protects her savings.

The fact is however, that neither moneylending companies nor the creditcooperative offer limited liability. Most loans had rigid repayment schedules,and significant additional costs in case of default. For example, it is commonfor moneylending companies to call the borrower, her guarantor or referenceperson, and/or their employers over the phone to demand payment for an un-

15

paid installment. Employers who receive these phone calls may fire the domes-tic worker, thus cutting off her income. The credit cooperative also does not limitthe borrower’s liability, in fact its loans are completely secured. It recovers un-paid loans by seizing the borrower’s and/or her guarantor’s savings. Short ofquitting the job, a worker cannot default on a loan from her employer since herpayments are deducted from her salary. If she did quit before she had paid backher loan she would likely find it impossible to find alternative employment inHong Kong.32 It thus seems implausible that MDWs borrow in order to avoidthe downside risk of their investment projects.

6.2 Short Time Horizon

Although repayment is enforced quite strongly within Hong Kong, moneylen-ders may be unable to enforce loan contracts once the worker leaves Hong Kong.If a migrant worker is uncertain about how much longer she will stay, this mayeffectively lower her cost of the loan.33

Migrants who have been re-employed by the same employer multiple timesmay be more secure about their job. This would be consistent with findinggreater indebtedness among migrant workers with a shorter tenure in their job.Instead, Figure 2 suggests that indebtedness rises as the number of years withthe current employer increases. This is likely connected to the fact that lendersreward workers who have a more secure job by offering them better loan terms.However it does not match the pattern we would expect if workers’ indebted-ness was solely caused by the insecurity of their jobs.

32Individuals who have breached their previous employment contract are unlikely to begranted a visa for a new contract[Government of the HKSAR, 2020].

33To our knowledge, there is no mechanism to prevent migrants from leaving Hong Kongwhile they are in debt. Moneylenders are aware of this risk, and partly mitigate this by schedul-ing loans to mature before the worker’s current employment contract ends. Thus to avoid re-paying the loan by leaving Hong Kong, the borrower would either have to run away, or hercontract would have to be terminated prematurely. Both employers and workers have the rightto terminate the contract at any time, with one month’s notice or one month’s payment in lieuof notice. Employers who terminate the contract are required to pay for the worker’s travel backto her home.

16

6.3 Loans as a Device to Solve Control Problems

We have shown about that individuals take loans even when they have suffi-cient savings. They then repay these loans from the monthly salaries they earnas domestic workers. This raises the following question. If instead they had dis-saved, then the same monthly salary could have been used to re-build the de-pleted savings. One reason individuals may not dissave is that it is difficult torebuild savings. Indeed, in informal interviews Filipino domestic workers agreethat it would be better to save than to borrow. However they report it is diffi-cult to save, because there is always a reason to spend the money instead. Thisself-reported inability to save has been documented in several other contextsas well. For example, it has provided a rationale for the success of simple sav-ings technologies in Kenya [Dupas and Robinson, 2013]. However, as we haveshown above, nearly all our subjects have at least one savings account; access tosavings products does not seem to facilitate their saving.

In turn, this suggests that loans may serve an additional purpose. The major-ity of the loans that we found in our survey data have strict repayment sched-ules, and so borrowers are committed to pay monthly installments until the loanis paid off. Since default carries large penalties, this could create a credible ra-tionale for avoiding other expenses. The prospect of default could help to resistthe temptation to buy oneself an unnecessary consumption good, or the pres-sure to purchase such items for one’s family or give gifts and treats to one’sfriends. Migrants who are sophisticated about their lack of control may thenactively choose to take a loan even if it were not financially necessary. Indeed,Baland, Guirkinger, and Mali [2011] argue that members of a credit cooperativein Cameroon borrow more than they need to, so that they can “pretend to bepoor”.

Roscas also help overcome the difficulty of saving. Gugerty [2007] reportsthat western Kenyan rosca members believe that the collective element givesthem the “strength to save”. In our context, roscas are not very common, proba-bly because roscas rely on mutual “trust”, which is more likely to develop whenmembers can monitor each other and enforce promises. These conditions are un-likely to develop organically in a population of transient urban migrants who

17

only meet once a week in a public location.34 Even among rosca members, wedo not find that roscas replace loans altogether, most likely because the roscapot is limited by the savings capacity of its members.35

In contrast, moneylending companies offer much larger loans. Access to theloans is easy, and to apply a domestic worker only needs to show her HongKong identity card and employment contract, and provide the phone numberof a reference person, or bring along a friend as a guarantor.36 The high interestrates and strict repayment schedules effectively limit future liquidity and flex-ibility to smooth consumption shocks, and in extreme situations can cause theworker to lose her income. Possibly these features actually make these loans at-tractive. Morduch [2010] discusses the case of a South Indian woman who tooka high-interest loan that she could have avoided. She believed the high interestrate incentivised her to pay back the loan much more quickly than she couldhave saved up the same amount.

7 Conclusion

Our research has benefitted from a large literature that precedes it. Manyscholars have noted that the poor do not save as much as they could[Banerjee and Duflo, 2011]. However others have also argued that borrowingremains an attractive choice for many poor individuals, even when they havethe wherewithal to save [Collins et al., 2009, Morduch, 2010]. This is because thehigh interest costs or the penalties for non-repayment induce the borrower to re-pay the loan and make it possible to avoid consumption, in a way that voluntarysavings mechanisms do not.

An important question in this context is: what are the compulsions that pre-

34Recall that 15% of our survey respondents belonged to roscas. The majority belonged to thesame island in the Philippines and knew each other well. However we found out informallythat two years after our study the rosca manager embezzled the pot; as of 2020 the members arestill waiting to get their savings back.

35Thirty percent of rosca members had a currently outstanding loan from a moneylender.Their mean loan size (HKD 25,500) was also similar to the mean loan size for those who did notbelong to roscas (HKD 23,149).

36Migrant domestic workers who either have a good repayment record or who have beenwith their current employer for longer than 5 years do not even need a reference person or aguarantor.

18

vent Filipino domestic workers in Hong Kong from saving successfully, but atthe same time allow them to repay loans regularly? A possible explanation isthe lack of self-control in the face of consumption opportunities. Certainly thereare abundant shopping opportunities in Hong Kong that might test an individ-ual’s self-control, and there are anecdotes about domestic workers who splurgeon consumption goods that might seem excessively expensive given their lowwages.

A second explanation points to the role of “kin taxes” or insecure propertyrights over one’s earnings and savings. Many of the Filipino workers we sur-veyed were earning considerably higher wages than their kin in the Philip-pines and were remitting money home regularly to support their expenses.Anderson and Baland [2011] argue that women participate in roscas in orderto wrest control from their husbands who might spend on unnecessary items.Ashraf, Aycinena, Martinez, and Yang [2015] find that El Salvadoran migrantsin the US deposited more into savings accounts when they had sole control overwithdrawals, than when they shared control with their relatives back home. Ina lab-in-the-field experiment Jakiela and Ozier [2016] find that Kenyan villageresidents are more unwilling to publicly reveal their earnings to a room fullof fellow residents, when a larger proportion of the fellow residents are theirkin. Baland et al. [2011] argue that Cameroonian credit cooperative memberstake loans instead of withdrawing their savings so that they can “pretend tobe poor” and avoid gifting or contributing to their friends and relatives. TheKenyan savers in the work by Dupas and Robinson [2013] also say their savingsboxes help them hide their money from their social network.

In informal interviews, Filipino domestic workers report that their familiesback home sometimes make unreasonable demands for money, and have un-realistically rosy ideas about their financial situation in Hong Kong. However,many also state that it is their responsibility to provide for their family, and is themain important reason why they are working in Hong Kong. Thus, although theliterature portrays the demands made by relatives as “taxes”, these relationshipscould be more complex in reality. Admittedly, financial support tends to flow inonly one direction from the migrant to her family members back home. How-ever the spouse, siblings, aunts and cousins in the Philippines are often looking

19

after the migrant’s children or elderly parents, overseeing house constructionand repair, or running the small business that the migrant has invested in. Thusthese may also be reciprocal arrangements, where one side provides financialsupport while the other provides manpower and facilitates peace of mind. Sim-ilarly, although friends in Hong Kong may borrow or request gifts, they alsolend in return and make gifts when the individual herself is in need. In futurework we will investigate the role that these social networks play in shaping theborrowing choices of Filipino domestic workers.

Ideally, possible interventions should be evaluated in the light of these possi-ble mechanisms. Commitment savings products that restrict the individual fromwithdrawing until a target date or savings balance is reached may be suitablefor individuals with present-biased time preferences [Ashraf, Karlan, and Yin,2006]. Those who wish to flexibly finance the expenses of their families may bebetter suited to a contractual savings product that requires them to replenishtheir savings after they have drawn them down [Morduch, 2010]. In a creditcooperative this could take the form of a combination loan-and-savings prod-uct where each month the borrower both repays the loan and makes a savingsdeposit. In future research we hope to investigate the effectiveness of these al-ternative products.

References

Siwan Anderson and Jean-Marie Baland. The economics of roscas and intra-household resource allocation. The Quarterly Journal of Economics, 117(3):963–995, 2011.

Nava Ashraf, Dean Karlan, and Wesley Yin. Tying oddyseus to the mast: Ev-idence from a commitment savings product in the philippines. The Quar-terly Journal of Economics, 121(2):635–672, 2006.

Nava Ashraf, Diego Aycinena, Claudia Martinez, and Dean Yang. Savings intransnational households: A field experiment among migrants from el sal-vador. The Review of Economics and Statistics, 97(2):332–351, 2015.

Jean-Marie Baland, Catherine Guirkinger, and Charlotte Mali. Pretending tobe poor: Borrowing to escape forced solidarity in cameroon. Economic De-velopment and Cultural Change, 60:1 – 16, 2011.

Abhijit Banerjee and Esther Duflo. Poor Economics. Random House, 2011.

20

Rashmi Barua, Gauri Kartini Shastry, and Dean Yang. Financial education forfemale foreign domestic workers in singapore. Economics of EducationReview, 2019.

Daryl Collins, Stuart Rutherford, Jonathan Morduch, and Orlanda Ruthven.Portfolios of the Poor: How the World’s Poor Live on $2 a Day. PrincetonUniversity Press, 2009.

Patricia Cortes and Jessica Pan. Outsourcing household production: Foreigndomestic workers and native labor supply in hong kong. Journal of LaborEconomics, 31(2):327–371, 2013.

Pascaline Dupas and Jonathan Robinson. Why don’t the poor save more? ev-idence from health savings experiments. American Economic Review, 103(4):1138–1171, 2013.

Government of the HKSAR. Foreign domestic helpers. https://www.immd.gov.hk/eng/services/visas/foreign_domestic_helpers.html, 2020. Immigration Department, Last accessed: 29Feb2020.

Mary Kay Gugerty. You can’t save alone: Commitment in rotating savings andcredit associations in kenya. Economic Development and Cultural Change,55(2):251–282, 2007.

Pamela Jakiela and Owen Ozier. Does africa need a rotten kin theorem? exper-imental evidence from village economies. Review of Economic Studies, 83:231–268, 2016.

Carolina Laureti. Why do poor people co-hold debt and liquid savings? Journalof Development Studies, 54(2):213–234, 2018.

Government of the Hong Kong SAR. Foreign domestic helpers and evolvingcare duties in hong kong. Research Brief, 2016-17(4), 2017. Office, Legisla-tive Council Secretariat.

PLUDWHK and Hong Kong Federation of Asian Domestic Workers. Betweena Rock and a Hard Place: The Charging of Illegal Agency Fees to FilipinoDomestic Workers in the Philippines and Hong Kong. 2016.

David McKenzie and Johan Mistiaen. Surveying migrant households: A com-parison of census-based, snowball and intercept point surveys. Journal ofthe Royal Statistical Society, 2:339–360, 2009. Series A: Statistics in Society.

Jonathan Morduch. Borrowing to save. Journal of Globalization and Develop-ment, 1:?, 2010.

Desmond Ng and Cheryl Tan. Borrowing, brokering, lend-ing: Inside the tangled web of maids and moneylenders.https://www.channelnewsasia.com/news/cnainsider/maids-domestic-workers-moonlight-brokers-moneylenders-borrowing-11304566,2019. 02 March.

21

Chi Wai Yu, Y. Jane Zhang, and Xuejing Zuo. Multiple switching and dataquality in the multiple price list. Review of Economics and Statistics, 2019.Forthcoming.

22

Figure 1: Average Monthly Savings Per Member in the Credit Cooperative ver-sus Previous Year’s Dividend Rate

020

4060

8010

0

1 1.5 2 2.5 3Dividend in previous year

Monthly net deposits per member versus dividend rate

23

Figure 2: Indebtedness versus Tenure with Current Employer

.1.2

.3.4

.5H

as a

n ou

tsta

ndin

g lo

an

0 10 20 30Years with current employer

Indebtedness as a function of job tenure

24

Table 1: Descriptive Statistics of Survey Respondents

Mean Standard error(1) (2)

Age 36.5 0.09Education

Less than high school 0.49 0.01High school 0.08 0.00Some post-high school 0.24 0.01Post-high school completed 0.19 0.01

Number of children 2.33 0.02Number of dependents 3.72 0.03Years since left Philippines 6.59 0.09Years in Hong Kong 4.84 0.08Years with current employer 3.26 0.06Mean salary (HKD per month) 4150.3 16.6Paid in cash 0.81 0.01Food provided 0.90 0.00Food allowance provided 0.10 0.00Remitted in past 2 months 0.88 0.00Mean remittances (HKD) 2163.5 20.2Percent of income remitted 0.52 0.00Remittance method

Bank 0.09 0.00Money service operator 0.64 0.01Online 0.01 0.00Other 0.17 0.01

Has bank account 0.83 0.01HKD account 0.32 0.01PHP account 0.91 0.00Has single-holder account 0.97 0.00Has joint account 0.06 0.00Total bank balance (HKD) 5839.4 206.5Mean savings per month (HKD) -39.8 18.7

Rosca member 0.11 0.00Uses a moneyguard 0.00 0.00Has outstanding debt 0.37 0.01If yes:

Total amount borrowed 21445.8 296.4Total repayment amount 27200.3 406.7Monthly repayment amount as % of salary 0.55 0.01

Has outstanding credit 0.14 0.03

Notes: The survey sample is reweighted to match the distribution of Filipino domesticworkers in the Hong Kong population, as estimated in the 5% micro-sample of the 2016Hong Kong By-census. 25

Tabl

e2:

Het

erog

eneo

usTr

eatm

entE

ffec

tsof

Hig

hR

ate

ofR

etur

non

Toke

nA

lloca

tion

Ris

k-av

ersi

onFi

nanc

iall

iter

acy

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Cha

ract

eris

tic

1.50

***

1.24

**1.

24**

0.14

0.72

-0.2

00.

03-3

.81

(0.5

1)(0

.52)

(0.5

2)(0

.84)

(2.0

5)(2

.09)

(2.1

1)(3

.31)

Hig

hin

tere

stra

te-2

.62

-21.

63*

-2.5

4-1

7.57

*(3

.38)

(11.

98)

(3.4

5)(1

0.60

)C

hara

cter

isti

c×

Hig

hin

tere

stra

te1.

77*

6.34

(1.0

7)(4

.23)

Age

0.19

0.17

0.22

0.25

0.24

0.23

(0.2

7)(0

.27)

(0.2

7)(0

.27)

(0.2

7)(0

.27)

Leng

thof

stay

-0.0

4-0

.02

-0.1

0-0

.04

-0.0

3-0

.01

(0.2

8)(0

.28)

(0.2

8)(0

.28)

(0.2

8)(0

.28)

Educ

atio

nH

igh

scho

ol-3

.26

-3.2

9-4

.40

-4.0

2-3

.98

-4.0

8(5

.95)

(5.9

6)(5

.98)

(6.0

4)(6

.05)

(6.0

4)So

me

post

-hig

hsc

hool

6.76

6.75

6.10

6.61

6.63

7.56

(4.8

5)(4

.86)

(4.8

6)(4

.90)

(4.9

1)(4

.94)

Com

plet

edpo

st-h

igh

scho

ol10

.07*

*9.

95**

9.55

**10

.70*

*10

.56*

*11

.30*

*(4

.81)

(4.8

2)(4

.81)

(4.8

5)(4

.85)

(4.8

7)C

onst

ant

35.1

8***

24.2

7**

26.3

7**

37.7

1***

49.7

1***

35.6

9***

37.1

1***

45.4

1***

(5.8

9)(1

1.09

)(1

1.43

)(1

3.30

)(5

.36)

(11.

87)

(12.

04)

(13.

23)

R-s

quar

ed0.

030.

070.

070.

080.

010.

050.

050.

06O

bser

vati

ons

324

324

324

324

324

324

324

324

Not

es:R

espo

nden

tsw

ere

rand

omly

assi

gned

toa

grou

pof

4-5

(say

,n)m

embe

rsan

dpl

ayed

nro

unds

wit

hth

atgr

oup

befo

rebe

ing

reas

-si

gned

toa

new

grou

p.T

heda

taco

ntai

nob

serv

atio

nsof

the

first

roun

dth

eypl

ayed

wit

hea

chgr

oup

they

wer

eas

sign

edto

.We

incl

ude

roun

ddu

mm

yva

riab

les

toco

ntro

lfor

roun

d-sp

ecifi

cef

fect

s.St

anda

rder

rors

inpa

rent

hese

s.**

*p<

0.01

,**p<

0.05

,*p<

0.1.

26

Table 3: Loan Characteristics

Overall Money-lender

Coop. Employer Friend /Relative

(1) (2) (3) (4) (5)

Fraction of loans 1.00 0.88 0.03 0.03 0.04

Principal 20453.5 22176.2 13153.9 15339.0 1480.8(272.8) (279.5) (182.44) (943.0) (159.98)

Repayment amount 25941.6 28461.0 13571.0 14254.2 1439.7(384.4) (396.52) (200.8) (1051.7) (131.7)

Interest rate 0.23 0.26 0.07 0.00 0.00(0.00) (0.00) (0.00) (0.00) (0.00)

Monthly schedule? 0.98 1.00 1.00 1.00 0.53(0.00) (0.00) (0.00) (0.00) (0.06)

Loan duration (months) 9.8 11.0 5.1 – –(0.12) (0.10) (0.20)

Notes: The survey sample is reweighted to match the distribution of Filipino domestic workers in theHong Kong population, as estimated in the 5% micro-sample of the 2016 Hong Kong By-census. Theannual inflation rate in Hong Kong in 2017 was 1.48%.

27

A1 Appendix

A1.1 Lab-in-the-Field Experiment

Each subject participated in a single experimental session each. Each sessionconsisted of 8-15 participants who sat at individual computer terminals. Eachsubject was randomly assigned to a group of 4-5 members, and played 4-5rounds of the decision-making experiment with this group before randomly be-ing assigned to a new group. At no point could they identify their group-matesamong the participants in the room.

In each round, participants were given an endowment of 100 tokens andasked to allocate them across three accounts (or “boxes”): the blue safe box thatwould give a certain return of x percent, the red box where, if the “investment”option was exercised the return would be 40% with probability 0.8 and zerootherwise; and the green box, that would generate a fixed in-kind return with aprobability proportional to the number of tokens placed in the box.

We experimentally varied the rate of return in the blue box to either be 3%or 10%. Experimental sessions were randomly assigned to one or the other rate.

The decision of whether to invest the amount in the red box was made by adifferent player in the group; we only analyse rounds where the participant wasnot investor. Also, to avoid endogenous token allocation in response to whatothers in the group did in previous periods, we only analyse the first round thatthe participant played with each group.

Each token placed in the green lottery box gave a 0.5% probability of suc-cess, so that if the participant placed 10 tokens in this box she would have a5% probability of winning a hand-bag as a prize. The total earnings from eachround were displayed to the player at the end of the round. At the end of thesession one round was randomly chosen and implemented, with an exchangerate of 1 token = $1. Thus the participant received the cash payment equal toher earnings, as well as the handbag, if she had won it in the randomly selectedround.

A1.2 Financial Literacy Questions

Question 1 Suppose you need to take a loan here in Hong Kong. There are twochoices. Loan A: You will get $10,000 for 6 months. You will have to payback $10,500 at the end of 6 months. Loan B: You will get $20,000 for 6months. You will have to pay back $20,800 at the end of 6 months.

Which loan is cheaper?

1

Which loan would you prefer?

Question 2 Suppose you need to take a loan of $10,000 here in Hong Kong.There are two choices. Loan A: You can get $10,000 for 6 months. You haveto repay $2,000 every month for 6 months. Loan B: You can get $10,000 for6 months. You have to repay $600 every week for 24 weeks.

Which loan is cheaper?

Which loan would you prefer?

A1.3 Re-weighting our Sample using the Random 5% Micro-sample from the 2016 Hong Kong By-Census

The 2016 Hong Kong Population By-Census sampled about one-tenth of all resi-dential quarters in Hong Kong, and collected detailed socio-economic data fromall households that lived there. We use the 5% sample of the micro-data releasedby the Census and Statistics Department, and consider the sub-sample of indi-viduals whose relationship to the household head is “live-in domestic helper”,nationality is Filipino and gender is female. We check that this sub-sample plau-sibly consists of Filipino domestic workers: all the individuals report that theyare currently working, their economic activity as “employees”, their industry as“domestic personnel” and occupation as one of the following three categories:“cleaners, helpers and related workers”, “personal care workers” and “driversand mobile machine operators”.

A simple comparison of the summary statistics for variables that are avail-able in both datasets suggests some differences in the age, education levels andlength of stay in Hong Kong. Accordingly, we construct the multivariate fre-quency distribution along these three dimensions in the Census dataset, andthen re-weight our survey sample accordingly.

Note that since the lab-in-the-field experiment implemented a randomizedintervention within the sample, un-weighted average and heterogenous treat-ment effects are internally valid.

2